Searcharxiv⌕ Search

arXiv subjects

Wei Wei

Publications and source records attributed to Wei Wei.

At least 37 records · Page 2Linked to original sources
arXiv2026

Deep Penalty Methods: A Class of Deep Learning Algorithms for Solving High Dimensional Optimal Stopping Problems

We propose a deep learning algorithm for high dimensional optimal stopping problems. Our method is inspired by the penalty method for solving free boundary PDEs. Within our approach, the penalized PDE is approximated using the Deep BSDE framework proposed by \cite{weinan2017deep}, which leads us to coin the term "Deep Penalty Method (DPM)" to refer to our algorithm. We show that the error of the DPM can be bounded by the loss function and $O(\frac{1}λ)+O(λh) +O(\sqrt{h})$, where $h$ is the step size in time and $λ$ is the penalty parameter. This finding emphasizes the need for careful consideration when selecting the penalization parameter and suggests that the discretization error converges at a rate of order $\frac{1}{2}$. We validate the efficacy of the DPM through numerical tests conducted on a high-dimensional optimal stopping model in the area of American option pricing. The numerical tests confirm both the accuracy and the computational efficiency of our proposed algorithm.

q-fin.MF↗
arXiv2026

CARB: A Covariate-Assessed Robust Borrowing Strategy with Literature-Informed Prior Weights for External Data

Borrowing external control data can improve the efficiency of clinical trials, particularly when patient accrual is difficult. A persistent challenge is how to prespecify the degree of borrowing systematically and transparently. In practice, prior weights are often selected heuristically or calibrated through simulation to achieve desired operating characteristics, making them difficult to justify scientifically. Moreover, patient-level covariate data from external sources are rarely available when the new trial is designed, precluding patient-level adjustment methods. We propose CARB (Covariate-Assessed Robust Borrowing), a framework that formalizes prior-weight specification as a design-stage assessment of baseline compatibility. Using only aggregate information, CARB quantifies discrepancies in prespecified baseline covariates between the new trial and each external source, without using outcome data from the new trial. A prespecified mapping translates the resulting dissimilarity measure into a source-specific prior weight on the exchangeable component of a robust borrowing model. Simulation studies show that CARB reduces bias and type I error inflation relative to fixed borrowing under observed and unobserved incompatibility, while improving efficiency when external controls are compatible. An application to advanced melanoma trials illustrates covariate-informed borrowing from multiple historical sources. An apparent discrepancy in reported baseline characteristics serves as a warning signal that reduces borrowing. CARB provides a transparent and reproducible way to borrow cautiously when patient-level covariate data from external sources are unavailable.

stat.ME↗
arXiv2026

Residual Flow Matching with Dynamic Cross-Interaction for 3D Multi-Person Motion Prediction

3D multi-person motion prediction requires modeling both individual kinematics and inter-person interactions. While Flow Matching is effective for multi-hypothesis generation to improve prediction accuracy, directly predicting skeletal sequences from pure noise often compromises structural consistency and introduces unreliable cross-agent interactions during early noise-dominated integration steps. To address this, we propose a Prior-Guided Residual Flow Matching framework. First, a Deterministic Coarse Prior (DCP) establishes a kinematic anchor, formulating the generative process as a conditional flow over motion residuals to simplify the generative objective and preserve structural stability. Second, a Dynamic Cross-Interaction (DCI) mechanism temporally synchronizes inter-agent message-passing with the integration progress, ensuring the extraction of reliable social contexts and improving multi-person motion fidelity. Finally, a decoupled joint-motion architecture with bidirectional fusion effectively preserves fine-grained kinematic coherence. Extensive experiments demonstrate that our approach achieves state-of-the-art prediction accuracy across multiple datasets. Code is available at https://github.com/Wei-Wei-a/Residual-Flow-Matching-with-Dynamic-Cross-Interaction-for-3D-Multi-Person-Motion-Prediction.

cs.CV↗
arXiv2026

AERIS: Offline Policy Improvement for Multi-UAV Integrated Sensing and Communication

Unmanned aerial vehicle (UAV)-enabled integrated sensing and communication (ISAC) is a promising 6G paradigm, but dynamic multi-UAV ISAC control must jointly balance communication quality, sensing reliability, and flight safety under stochastic mobility. Existing optimization methods often require repeated global non-convex solving, while online reinforcement learning (RL) depends on risky trial-and-error flights that may cause sensing loss or collision-risk events. This paper proposes AERIS, an offline policy improvement framework for multi-UAV ISAC. AERIS learns from fixed flight logs under centralized training and decentralized execution, so each UAV acts from local histories while training uses logged global information to assess team-level effects. We further design STAR-CRDT, an offline multi-agent RL algorithm that performs support-aware local action rectification and distills only trusted improvements into the decentralized actor. We prove an offline-support policy improvement guarantee. Experiments show that STAR-CRDT improves the main ISAC objective return by 29.3% over the strongest baseline. It further improves communication sum rate, sensing pass rate, and sensing margin by 3.4%, 4.8%, and 69.1%, while reducing collision-risk events by 54.2%. On unseen real-road maps built from OpenStreetMap data, STAR-CRDT still obtains the best return.

cs.NI↗
arXiv2026

WarpSAC: Towards the Pinnacle of Scalable Off-policy RL by Rethinking Exploration and Exploitation

Massively parallel simulation changes the data regime in which off-policy reinforcement learning (RL) is trained, challenging stabilizers designed for data-limited replay. Through controlled experiments across eight benchmark families, we show that these stabilizers are data-regime-dependent: parameter normalization helps with narrow replay coverage but restricts value fitting when data are abundant, while clipped double-Q can be relaxed in high-throughput manipulation. Age-biased replay weighting improves learning efficiency across regimes, especially with limited network capacity. Based on these findings, we propose WarpSAC, a regime-aware family of off-policy RL algorithms. WarpSAC uses Sample Weight Decay for efficient exploitation and provides two variants: WarpSAC-L (Norm ON, clipped double-Q) for data-limited CPU-scale training, and WarpSAC-A (Norm OFF, single-Q) for data-abundant GPU-parallel training. WarpSAC improves normalized score--step AUC over FlashSAC by 4.5% across nine CPU-scale environments and 23.1% across fourteen GPU-parallel environments. It increases UnitreeG1TransportBox-v1 success rate from 19.8% to 96.4%, improves mean normalized wall-time AUC on MuJoCo Playground by 19.1%, and achieves 36.4% faster sim-to-real deployment on Unitree G1 than FlashSAC. These results show that scalable off-policy RL should adapt its stabilizers to the available data regime.

cs.LG↗
arXiv2026

MACD: Multi-Agent Clinical Diagnosis with Self-Learned Knowledge for LLM

Large language models (LLMs) have shown promise in supporting medical diagnosis, with prompting-based methods offering a flexible and deployable means of capability enhancement. However, existing prompt engineering and multi-agent approaches often focus on optimizing single inferences, paying less attention to the accumulation of reusable experience from clinical practice, constraining their real-world applicability. To address this, this study proposes a novel Multi-Agent Clinical Diagnosis (MACD) framework, which allows LLMs to self-learn clinical knowledge via a multi-agent pipeline that summarizes, refines, and applies diagnostic insights, mirroring the professional development of human physicians. We further extend it to a MACD-human collaborative workflow, where multiple LLM-based diagnostician agents engage in iterative consultations, supported by a judge agent and human oversight for cases where agreement is not reached. The MIMIC-MACD cohort comprising 4,390 real-world patient cases across seven diseases is constructed, including 1,314 cases for knowledge learning and 3,076 held-out cases for evaluation. Across diverse open-weight LLMs, MACD significantly improves primary diagnostic accuracy, achieving an average improvement of 11.6 percentage points over established authoritative knowledge, while narrowing the performance gap between open-weight models and state-of-the-art LLMs. Furthermore, the MACD-human workflow yields an 18.3-percentage-point improvement over physician-only diagnosis on text-only vignettes, demonstrating the synergistic potential of human-AI collaboration. This work thus presents a scalable self-learning paradigm that bridges the gap between the intrinsic knowledge of LLMs and the demands of real-world clinical practice, advancing towards a reliable, interpretable, and deployable AI-assisted diagnosis.

cs.AI↗
arXiv2026

LLM-based Agents for Forecasting and Prediction: Methods, Training, Evaluation, and Applications

Large language models (LLMs) now support forecasting systems that combine language-based reasoning with temporal data, evidence retrieval, external tools, and iterative prediction. We investigate LLM-based forecasting agents, meaning systems in which a language model contributes to a scored prediction about a future or currently unobserved target. We organize architectures into three groups. Standalone LLM workflows operate on encoded time series or event context. Tool- and retrieval-augmented agents incorporate external evidence. Hybrid systems pair LLMs with statistical or foundation models. We then review training methods and evaluation protocols. We examine negative as well as positive evidence, including sensitivity to small input perturbations, ablations in which the LLM component does not improve accuracy, and benchmark gains that may reflect contamination instead of temporal reasoning. We cover applications in finance, weather, health, energy, and operations, and we summarize the benchmarks and datasets used for evaluation. The evidence indicates that measurement is a central limitation. Future work requires calibration under distribution shift, contamination-resistant live evaluation, explicit reporting of cost and accuracy together, and methods for handling feedback between deployed forecasts and the outcomes being forecast.

cs.AI↗
arXiv2026

HeatTok: Enhancing Remote Sensing Image Understanding via Thermodiffusion-based Tokenization

Current visual tokenizers in Multimodal Large Language Models (MLLMs) predominantly rely on patch-based partitioning, which causes severe semantic mixture and object fragmentation in remote sensing imagery due to the irregular contours of geo-objects. Moreover, existing adaptive methods struggle to extract precise object-level tokens and lack dedicated geometric positional encodings for irregular regions. In this paper, we propose HeatTok, a semantic-aware tokenizer driven by thermodiffusion aggregation. Inspired by the physical principles of heat conduction, HeatTok adaptively merges adjacent homogeneous regions to generate semantically independent, object-aligned irregular tokens. To enable MLLMs to perceive these irregular shapes, we design the Gaussian Multimodal Rotary Positional Embedding (G-MRoPE), which models token spatial distributions via 2D Gaussians and explicitly injects center, scale, and orientation cues. Extensive evaluations on the VRSBench and EarthVQA datasets demonstrate that HeatTok effectively preserves object-level semantic integrity and achieves state-of-the-art performance under a reasonable token budget. The code is available: https://github.com/YingyingYan1/HeatTok.

cs.CV↗
arXiv2026

Liouville theorem for a class of p-Laplace type equations on manifolds

We study a class of $p$-Laplace equations $$Δ_p u-λu^{p-1}+ u^{q-1}=0$$ on a closed $n$-dimensional Riemannian manifold $(M,g)$ with $\operatorname{Ric}\geqslant(n-1)g$. For $1 2$ and $p 0$; aside from the constant solution, the equation admits a positive nonconstant solution. This answers Véron's problem raised in \cite{Ver92}.

math.AP↗
arXiv2026

Efficient Resource Optimization for Split Federated Learning

Split federated learning (SFL) has emerged as a powerful paradigm for model training at the edge. However, SFL inherently involves discrete decision variables for model splitting and resource allocation, resulting in a challenging mixed-integer problem. Consequently, prior optimization schemes for SFL are either \textit{heuristic} or \textit{computationally inefficient}, which cannot handle large-scale user populations. To address this limitation, this work establishes an efficient optimization framework for SFL under resource-constrained networks. Our framework jointly optimizes model splitting and resource allocation to minimize training cost, which is defined as the weighted sum of latency and energy costs. We first study the model splitting problem and develop a polynomial-time algorithm that achieves the global optimum. Then, we extend the approach to the joint model splitting and resource allocation problem. In this case, we formulate it as a two-dimensional master problem and develop an efficient approximation method with a $(1+ε)$-approximation guarantee. Extensive experiments show that the proposed approach provides efficient solutions to strike the optimal energy--latency tradeoff.

cs.LG↗
arXiv2026

CoCA: Step-level Reward for Free in RL-based T2I Diffusion Model Fine-tuning

Recent advances in text-to-image (T2I) diffusion model fine-tuning leverage reinforcement learning (RL) to align generated images with learnable reward functions. The existing approaches reformulate denoising as a Markov decision process for RL-driven optimization. However, they suffer from reward sparsity, receiving only a single delayed reward per generated trajectory. This flaw hinders precise step-level attribution of denoising actions, undermines training efficiency. To address this, we propose a simple yet effective credit assignment framework that dynamically distributes dense rewards across denoising steps. Specifically, we track changes in cosine similarity between intermediate and final images to quantify each step's contribution on progressively reducing the distance to the final image. Our approach avoids additional auxiliary neural networks for step-level preference modeling and instead uses reward shaping to highlight denoising phases that have a greater impact on image quality. Our method achieves 1.25 to 2 times higher sample efficiency and better generalization across four human preference reward functions, without compromising the original optimal policy.

cs.CV↗
arXiv2026

Epistemic Tensions: Reframing A Visualization Co-Design through Entanglement Theory

In this work, we present how employing the lens of entanglement helped us examine and reframe epistemic tensions arising in a visualization co-design project. Entanglement theory challenges traditional assumptions in the visualization research community by emphasizing that knowledge is not produced through linear, isolated processes, but is inherently entangled with phenomena and apparatuses. While this perspective offers a compelling critique of conventional research practices, its practical value for visualization research remains underexplored. We apply the entanglement lens to examine and reframe the epistemic tensions that emerged in a longitudinal community-based visualization co-design project. Our experience shows that the entanglement perspective not only provides a richer understanding of these tensions, but also helps transform them into generative opportunities for methodological and theoretical reflection. Applying this lens enabled us to critically interrogate the language used in research, to develop a more nuanced understanding of visualization co-design, and to surface ``dark sides'' of conventional visualization design pipelines. These contributions illustrate the practical value of embracing entanglement as an epistemological lens for visualization research.

cs.HC↗
arXiv2026

Multi Interests for Joint Search-Recommendation Modeling

Search and recommendation are crucial for understanding user preferences. More and more studies are attempting to jointly model search behavior and recommendation behavior, by integrating user active search and passive recommendation behavior data to better mine user preferences. However, although existing cross-domain unified modeling frameworks can effectively compensate for the differences in behavior between domains, they overlook the expression of interests in different scenarios under mixed sequences. In this study, we propose a multi-interest-based mixed sequential modeling framework MIJSR, which performs multi-interest mining and adaptive integration on search recommendation mixed sequences from both structural and semantic perspectives. Specifically, our model can be roughly divided into three modules: cross-domain behavior fusion, multi-interest mining, and multi-task prediction. Firstly, we align the representations of query and item through contrastive learning training. Then, we extract the multi interests of the mixed behavior sequence from both structural and semantic perspectives. Structurally, we extract search interests, recommendation interests, and cross interests through subsequence partitioning and mask settings; In terms of semantics, we use the semantic information of queries for clustering and perform semantic segmentation on mixed sequences to construct semantic multi interests. Finally, the adaptive fusion of multiple interests is combined with other side information to use a progressive layered extraction model for multi-task prediction. Extensive experiments on two open-source datasets have shown that our model can further enhance its accuracy in search and recommendation by extracting users' multi interests at a fine-grained level. Codes are available at https://github.com/pxcstart/MIJSR.

cs.IR↗
arXiv2026

Market-Information-Aware Gated-LoRA of Foundation Models for Transferable Day-Ahead Electricity Price Forecasting

Electricity price forecasting is crucial for market participants but remains difficult because prices are volatile, market-specific, and closely tied to anticipated system conditions. Existing supervised methods depend largely on market-specific historical data, limiting their use in newly established or data-scarce markets. This paper proposes a market-information-aware adaptation framework that transfers the Chronos-2 time-series foundation model to day-ahead electricity price forecasting. It first constructs a multi-source market information (MSMI) interface aligning 7-day price context with pre-clearing supply--demand, reserve, maintenance, generator-capacity, and intertie variables, and then trains a source-domain gated low-rank adapter (LoRA), updating about $1\%$ of model parameters without target-market labels. The gate scales the frozen source adapter according to reserve-tightness and operating-state signals. A leave-one-market-out protocol is adopted for evaluating cross-market transferability. Experiments on four Chinese provincial day-ahead spot markets show that the proposed framework reduces the average MAE/RMSE by $6.24\%/7.99\%$ relative to market-information-aware zero-shot Chronos-2 and by $3.05\%/3.52\%$ relative to vanilla Source-LoRA. Experiments show that the gain is not reproduced by a learned global scalar or by random gate initialization, while the additional improvement over Source-LoRA is limited. These results suggest that market-structured inputs and state-dependent gated LoRA can provide a practical transfer path for data-scarce electricity markets.

cs.LG↗
arXiv2026

Remember-R1: Mitigating Long-Context Visual Forgetting through Reinforcement Learning

Multimodal large language models (MLLMs) increasingly rely on long chain-of-thought reasoning for complex tasks. However, as reasoning sequences lengthen, models may gradually rely less on visual evidence and more on accumulated textual context, leading to visual forgetting. Existing approaches do not directly constrain how visual evidence is used and maintained along the original reasoning trajectory, leaving long-context visual forgetting insufficiently addressed. To address this issue, we propose Remember-R1, a reinforcement learning framework that mitigates long-context visual forgetting by applying process-level supervision directly on the original reasoning trajectory. Specifically, Remember-R1 introduces rewards that encourage broader coverage of matched visual keywords, stronger persistence of visual dependence in later reasoning steps, and greater focus on question-relevant image regions. Experiments across multiple model scales and diverse multimodal benchmarks demonstrate that Remember-R1 consistently improves reasoning performance. Additional analyses further show that it slows the decline of visual attention during generation, supporting its effectiveness in mitigating long-context visual forgetting.

cs.CV↗
arXiv2026

What Does Chain-of-Thought Contribute at Probe Time? Evidence for Local Co-Occurrence Activation

Chain-of-thought (CoT) prompting enhances large language model performance, yet what drives these gains remains unclear. We study this question from a probe-time perspective: holding CoT rationales fixed, we test which textual properties matter for the final prediction. Across multiple datasets and model configurations, we find that randomizing the order of rationale sentences has little effect on accuracy, suggesting that the global order of reasoning steps is not the main source of the probe-time benefit. Moreover, even when the words in a rationale are randomly reordered, performance remains well above the no-rationale baseline, indicating that the rationale's words remain useful even without their original order. Restoring only short-range word order further improves performance and brings it substantially closer to full CoT. In most settings, much of this local-order gain is already obtained with three-word windows. Control experiments rule out explicit answer copying, simple lexical cues, generic topical context, and general robustness to shuffling as the main explanations. Mechanistic analyses further show that short-window gains are largely formed in early-to-middle model layers, with answer-relevant evidence concentrated in local text spans. Together, these findings support a local co-occurrence activation (LCA) interpretation: the probe-time benefit of fixed rationales arises mainly from the words they contain and short-range word co-occurrences.

cs.AI↗
arXiv2026

GaussianSeed: Hierarchical Gaussian Seeding for High-Resolution 3D Occupancy Prediction

Vision-centric 3D occupancy prediction provides dense scene representations essential for autonomous driving and robotic navigation, yet existing methods struggle to scale to high voxel resolutions due to prohibitive computational costs. To address this, we introduce GaussianSeed, a progressive multi-scale Gaussian occupancy prediction framework that organizes primitives into a coarse-to-fine hierarchy. Benefiting from this hierarchical design, GaussianSeed effectively circumvents the memory bottlenecks inherent in dense representations, successfully scaling to a $0.1\text{m}$ spatial resolution while maintaining real-time inference capabilities. To comprehensively evaluate high-resolution geometric perception, we further construct TJScenes, a panoramic six-camera occupancy dataset with highly detailed $0.1\text{m}$ annotations. Extensive experiments on Occ3D-nuScenes and TJScenes demonstrate that GaussianSeed delivers the lowest latency among all evaluated methods while maintaining highly competitive accuracy, advancing the efficiency-quality frontier of high-resolution 3D occupancy prediction. Codes are available at https://github.com/Athameral/GUSD

cs.CV↗
arXiv2026

The Exact Maximum of the Spectral Sum of Graphs

For a simple graph $G$ of order $n$, let $S_2(G)=λ_1(G)+λ_2(G)$ denote its spectral sum. We determine, for every $n\geq5$, the exact maximum of $S_2(G)$ and all equality cases. The unique maximizer, up to isomorphism, is the complement of the disjoint union of a suitably balanced complete bipartite graph and isolated vertices, with the sizes of its three parts determined by $n$ modulo $7$. Denoting this graph by $K_n^\star$, we further show that $ S_2(K_n^\star)\leq\frac{8n}{7}-2,$ with equality exactly when $7\mid n$. This proves a conjecture of Kumar, Liu, Monterde, Pragada and Tait, which strengthens the Aouchiche--Hansen 2010 conjecture by extending it from connected graphs to all graphs and by asserting uniqueness of the extremal graph. The result also subsumes the 2008 conjecture of Ebrahimi B., Mohar, Nikiforov, and Ahmady. The proof combines Ky Fan's variational principle with a spectral inequality for weighted Ferrers quotients to reduce the problem to an explicit family whose complements have incidence rank one. Exact integer optimization and a separate equality analysis then yield the maximum and uniqueness.

math.CO↗